In the rapidly evolving field of quantum computing, silicon spin qubits are emerging as a leading candidate for building scalable, fault-tolerant quantum computers. A new review titled "Single-Electron Spin Qubits in Silicon for Quantum" published May 2 in Intelligent Computing , a Science Partner Journal, highlights the latest advances, challenges and future prospects of silicon spin qubits for quantum computing.
Silicon spin qubits are compatible with existing semiconductor industry manufacturing processes, making them promising for universal quantum computers. They have several remarkable properties. "They can have long coherence times, up to 0.5 seconds, single-qubit gate fidelities exceeding 99.95%, and two-qubit gate fidelities surpassing the fault-tolerant threshold," according to the authors. In addition, silicon spin qubits can operate as "hot qubits" at temperatures of 1 Kelvin or above, and recent studies have even demonstrated gate fidelities required for fault-tolerant operations at this temperature.
Silicon quantum dots, also called artificial atoms, are the basic structure of silicon spin qubits. These tiny structures can trap and control individual electrons, allowing researchers to define various types of spin qubits. For example, single-electron dots can be manipulated with alternating-current magnetic fields, while two-electron systems in double dots use exchange interactions to define qubits, such as singlet-triplet qubits, and to construct various two-qubit gates, including SWAP gates, controlled-phase gates, and controlled-not gates.
The review focuses on two main types of silicon spin qubits: gate-defined quantum dots and donor-based quantum dots. Gate-defined quantum dots are built using electric fields to trap electrons, with fabrication relying on substrates like silicon, silicon/germanium heterostructures, or silicon metal-oxide-semiconductor structures. Whilst donor-based quantum dots take a different approach, encode qubits through the use of dopants like phosphorus, with fabrication methods including ion implantation and scanning tunneling microscope lithography.
However, the gate-defined quantum dots and the donor-based quantum dots share common technologies. Their spin coherence times for both quantum dots are significantly prolonged using isotopically purified materials. Their qubit initialization and readout can be achieved through spin-to-charge conversion processes, such as spin-selective tunneling and Pauli spin blockade. Single-qubit gates can be manipulated using electron spin resonance or electric dipole spin resonance techniques. Two-qubit gates are implemented by utilizing the exchange interaction between qubits.
Realizing the long-distance coupling of spin qubits is crucial for increasing the number of qubits and enabling distributed quantum computing architectures. Circuit quantum electrodynamics, which uses microwave photons in superconducting resonators, are making this possible. Strong spin-photon coupling has been demonstrated using hybrid techniques like synthetic spin-orbit interactions provided by micromagnets. These advances allow for coherent quantum state transfer between distant qubits, supporting the development of quantum multi-core processors and distributed architectures.
The future of silicon spin qubits looks promising but also faces challenges. For gate-defined quantum dots, "integrating silicon qubits with on-chip classical control, exploring new 2D and 3D qubit array layouts, and possibly operating at higher temperatures are important research areas," according to the authors. For donor-based quantum dots, further development of fabrication techniques, integration with "hot qubits" and cryogenic complementary metal-oxide semiconductor technology, and exploring new dopants are areas of focus. Scaling up will require continuous improvement in qubit operation fidelity, addressing inhomogeneity and disorder in large-scale arrays, and optimizing the architecture.
